Automation in sleep state classification and maturity evaluation of EEG in preterm-born infants
6 affiliations
- 1Graduate School of Education, The University of Tokyo, Bunkyo, Tokyo, Japan; Division of Neural Dynamics, National Institute for Physiological Sciences, Okazaki, Aichi, Japan; Correspondence to: Division of Neural Dynamics, National Institute for Physiological Sciences, 38 Nishigonaka, Myodaiji-cho, Okazaki-shi, Aichi 444-8585, Japan.
- 2Graduate School of Education, The University of Tokyo, Bunkyo, Tokyo, Japan
- 3Department of Pediatrics, Nagoya University Graduate School of Medicine, Nagoya, Aichi, Japan
- 4Department of Obstetrics and Gynecology, Nagoya University Graduate School of Medicine, Nagoya, Aichi, Japan
- 5Division of Neonatology, Center for Maternal-Neonatal Care, Nagoya University Hospital, Nagoya, Aichi, Japan
- 6Department of Pediatrics, Nagoya University Graduate School of Medicine, Nagoya, Aichi, Japan; Department of Developmental Disability Medicine, Nagoya University Graduate School of Medicine, Nagoya, Aichi, Japan
Abstract
There is a growing need to automate EEG monitoring of preterm-born infants, particularly where expert evaluation is limited. However, existing models necessitate human intervention for comprehensive assessment, thereby limiting the clinical applicability. We collected 165 EEG records from 65 preterm-born infants. 8-channel EEG records were segmented every 30 s and allocated into eight classes (three preterm-equivalent, four term-equivalent, and one noise class). We performed noise detection, sleep state classification, and maturity assessment through the classification. ImageNet pretrained ResNet-50 was finetuned to classify the segmented EEG images. A novel index, named “maturity rate,” was calculated as the proportion of term-equivalent labels in each EEG recording. The model achieved a macro F1 score of 0.58 and Cohen’s Kappa of 0.58 for the 8-class classification. When the labels were merged into active sleep, quiet sleep, and noise categories, performance improved to an F1 score of 0.84 and Kappa of 0.74. The maturity rate increased continuously with postmenstrual age at the time of recording, reflecting functional brain development beyond the term–preterm dichotomy. Our model enables end-to-end EEG assessment of sleeping newborns by simultaneously performing noise detection, sleep state classification, and maturity assessment, with minimal need for human intervention.
Related
- Altered connectivity and network efficiency in preterm infants
- EEG-based deep learning estimates preterm infants’ cerebral age: Contribution of slow-wave and cross-frequency coupling
- Distinct experience-dependent reorganization of rhythmic syllable tracking in 6-month-old term and preterm infants
- Early evolution of oral feeding skills in very preterm infants: impact of an oral stimulation protocol assessed by cot-side suckometer: a pilot study
- Transcriptomics-guided serum podoplanin and interpretable machine learning for early risk stratification of necrotizing enterocolitis in preterm infants
BibTeX
@article{fujihira2026automation,
title = {{Automation in sleep state classification and maturity evaluation of EEG in preterm-born infants}},
author = {Ryo Fujihira and Hama Watanabe and Anna Shiraki and Yuji Ito and Takafumi Ushida and Hiroyuki Yamamoto and Tomohiko Nakata and Yoshiaki Sato and Jun Natsume and Hiroyuki Kidokoro and Gentaro Taga},
journal = {Neuroscience Research},
year = {2026},
month = oct,
doi = {10.1016/j.neures.2026.105139},
url = {https://doi.org/10.1016/j.neures.2026.105139},
note = {CC BY},
}
RIS
TY - JOUR
TI - Automation in sleep state classification and maturity evaluation of EEG in preterm-born infants
AU - Ryo Fujihira
AU - Hama Watanabe
AU - Anna Shiraki
AU - Yuji Ito
AU - Takafumi Ushida
AU - Hiroyuki Yamamoto
AU - Tomohiko Nakata
AU - Yoshiaki Sato
AU - Jun Natsume
AU - Hiroyuki Kidokoro
AU - Gentaro Taga
PY - 2026
DA - 2026/10/30
JO - Neuroscience Research
DO - 10.1016/j.neures.2026.105139
UR - https://doi.org/10.1016/j.neures.2026.105139
AB - There is a growing need to automate EEG monitoring of preterm-born infants, particularly where expert evaluation is limited. However, existing models necessitate human intervention for comprehensive assessment, thereby limiting the clinical applicability. We collected 165 EEG records from 65 preterm-born infants. 8-channel EEG records were segmented every 30 s and allocated into eight classes (three preterm-equivalent, four term-equivalent, and one noise class). We performed noise detection, sleep state classification, and maturity assessment through the classification. ImageNet pretrained ResNet-50 was finetuned to classify the segmented EEG images. A novel index, named “maturity rate,” was calculated as the proportion of term-equivalent labels in each EEG recording. The model achieved a macro F1 score of 0.58 and Cohen’s Kappa of 0.58 for the 8-class classification. When the labels were merged into active sleep, quiet sleep, and noise categories, performance improved to an F1 score of 0.84 and Kappa of 0.74. The maturity rate increased continuously with postmenstrual age at the time of recording, reflecting functional brain development beyond the term–preterm dichotomy. Our model enables end-to-end EEG assessment of sleeping newborns by simultaneously performing noise detection, sleep state classification, and maturity assessment, with minimal need for human intervention.
SN - 0168-0102
SN - 1872-8111
ER -